train bc_transformer

[Nimingez]

I collect my wipe datase from robosuite

python dataset_states_to_obs.py --dataset /home/xxx/robomimic/demo.hdf5 --output_name image_dense.hdf5 --done_mode 2 --camera_names agentview robot0_eye_in_hand --camera_height 84 --camera_width 84 I want to train with bc_transformer

python train.py --config ../config/default_templates/bc_transformer.json --dataset /home/xxx/robomimic/image.hdf5

RuntimeError: cannot reshape tensor of 0 elements into shape [-1, 0] because the unspecified dimension size -1 can be any value and is ambiguous

[amandlek]

The default config trains on object states not image observations (see here) - is this what you wanted?

As for the error, my best guess is that there is something funny about the "object" key for the Wipe task (populated here during dataset_states_to_obs.py) so it might be worth checking the shapes going into this observation key. This might actually be a robosuite issue (with the Wipe task).

Let us know what you find and we can follow up.

[BuiDacHien]

Hey sir, trong lúc tôi chạy đoạn code sau:

python robomimic/scripts/train.py --config robomimic/exps/templates/bc.json --dataset datasets/lift/ph/low_dim_v141.hdf5 --debug

Tôi gặp lỗi sau:

100%|##########| 3/3 [00:00<00:00, 8.87it/s] Train Epoch 1 { "Cosine_Loss": 0.5587675174077352, "L1_Loss": 0.0957380086183548, "L2_Loss": 0.19192640483379364, "Loss": 0.19192640483379364, "Optimizer/policy0_lr": 0.0001, "Policy_Grad_Norms": 0.2235456732159946, "Time_Data_Loading": 0.00018465916315714518, "Time_Epoch": 0.005668433507283529, "Time_Log_Info": 1.4901161193847656e-05, "Time_Process_Batch": 0.00012256701787312826, "Time_Train_Batch": 0.005309998989105225 } video writes to /tmp/tmp_trained_models/test/20240310193926/videos/Lift_epoch_1.mp4 rollout: env=Lift, horizon=10, use_goals=False, num_episodes=2 100%|##########| 2/2 [00:18<00:00, 9.12s/it]

Epoch 1 Rollouts took 9.12014651298523s (avg) with results: Env: Lift { "Horizon": 10.0, "Return": 0.0, "Success_Rate": 0.0, "Time_Episode": 0.3040048837661743, "time": 9.12014651298523 } save checkpoint to /tmp/tmp_trained_models/test/20240310193926/models/model_epoch_1_Lift_success_0.0.pth

Epoch 1 Memory Usage: 1238 MB

100%|##########| 3/3 [00:00<00:00, 51.14it/s] Train Epoch 2 { "Cosine_Loss": 0.5759696165720621, "L1_Loss": 0.0873916173974673, "L2_Loss": 0.1788101146618525, "Loss": 0.1788101146618525, "Optimizer/policy0_lr": 0.0001, "Policy_Grad_Norms": 0.022320906980591342, "Time_Data_Loading": 0.0001711289087931315, "Time_Epoch": 0.0009937047958374023, "Time_Log_Info": 1.8628438313802085e-05, "Time_Process_Batch": 1.510779062906901e-05, "Time_Train_Batch": 0.0007715781529744466 } video writes to /tmp/tmp_trained_models/test/20240310193926/videos/Lift_epoch_2.mp4 rollout: env=Lift, horizon=10, use_goals=False, num_episodes=2 100%|##########| 2/2 [00:12<00:00, 6.23s/it]

Epoch 2 Rollouts took 6.228485345840454s (avg) with results: Env: Lift { "Horizon": 10.0, "Return": 0.0, "Success_Rate": 0.0, "Time_Episode": 0.2076161781946818, "time": 6.228485345840454 }

Epoch 2 Memory Usage: 1256 MB

finished run successfully! Exception ignored in: <function MjRenderContext.del at 0x7f027bf3d310> Traceback (most recent call last): File "/dis/camvan/BDH/robosuite/robosuite/utils/binding_utils.py", line 199, in del self.gl_ctx.free() File "/dis/camvan/BDH/robosuite/robosuite/renderers/context/egl_context.py", line 149, in free EGL.eglMakeCurrent(EGL_DISPLAY, EGL.EGL_NO_SURFACE, EGL.EGL_NO_SURFACE, EGL.EGL_NO_CONTEXT) File "/home/camvan/anaconda3/envs/robomimic_venv/lib/python3.8/site-packages/OpenGL/error.py", line 230, in glCheckError raise self._errorClass( OpenGL.raw.EGL._errors.EGLError: EGLError( err = EGL_NOT_INITIALIZED, baseOperation = eglMakeCurrent, cArguments = ( <OpenGL._opaque.EGLDisplay_pointer object at 0x7f022bcc19c0>, <OpenGL._opaque.EGLSurface_pointer object at 0x7f028456a040>, <OpenGL._opaque.EGLSurface_pointer object at 0x7f028456a040>, <OpenGL._opaque.EGLContext_pointer object at 0x7f0284661f40>, ), result = 0 ) Exception ignored in: <function EGLGLContext.del at 0x7f027bf3d160> Traceback (most recent call last): File "/dis/camvan/BDH/robosuite/robosuite/renderers/context/egl_context.py", line 155, in del self.free() File "/dis/camvan/BDH/robosuite/robosuite/renderers/context/egl_context.py", line 149, in free EGL.eglMakeCurrent(EGL_DISPLAY, EGL.EGL_NO_SURFACE, EGL.EGL_NO_SURFACE, EGL.EGL_NO_CONTEXT) File "/home/camvan/anaconda3/envs/robomimic_venv/lib/python3.8/site-packages/OpenGL/error.py", line 230, in glCheckError raise self._errorClass( OpenGL.raw.EGL._errors.EGLError: EGLError( err = EGL_NOT_INITIALIZED, baseOperation = eglMakeCurrent, cArguments = ( <OpenGL._opaque.EGLDisplay_pointer object at 0x7f022bcc19c0>, <OpenGL._opaque.EGLSurface_pointer object at 0x7f028456a040>, <OpenGL._opaque.EGLSurface_pointer object at 0x7f028456a040>, <OpenGL._opaque.EGLContext_pointer object at 0x7f0284661f40>, ), result = 0 )